Hemophilia: New Protein Therapeutics

The recurrent hemarthroses and resultant crippling arthropathy of hemophilia have almost been completely eradicated through aggressive prophylactic therapy with clotting concentrates in the developed world when initiated at a young age. Quality-of-life measures in children now track similar to their unaffected peers, and patients engage regularly in sporting activities unimagined in previous generations.¹  Prophylaxis for severe hemophilia is now recognized as the standard of care with optimal initiation very early in life before the onset of repeated hemarthroses, typically between ages 1 to 3 years.²  Nevertheless, barriers remain to realizing the benefits of prophylaxis universally. The development of alloantibodies that inhibit the activity of infused replacement products remains a significant complication (up to 30% of patients with hemophilia A). The costs for prophylactic replacement therapy are much greater than $100,000 per patient per year. Repeated venous access (typically three times per week to every other day) is still a barrier for many. Suboptimal adherence to a prophylactic regimen also compromises outcomes.³  There is also the frequent requirement for central venous access devices in the youngest boys to facilitate prophylactic infusion with associated risk for infection and thrombosis.⁴  Despite the increased capacity of factor concentrate production, worldwide demand for factor VIII is now over 5 billion units per year, most of which is infused in North America and Europe, with 80% of the world still without proper access to replacement therapy. Recombinant DNA technology is now serving as a platform for continued innovation to try to overcome some of the remaining hurdles for hemophilia care. This article summarizes the current and future trends in hemophilia therapeutics, with a particular focus on targeted bioengineering strategies.

Purified plasma-derived (pd) factor VIII (FVIII) concentrates have been available since the 1970s, but require rigorous plasma donor screening and viral inactivation technologies to reduce the risk for blood-transmitted pathogens. Recombinant FVIII (rFVIII) concentrates, available since 1992, are derived from expression within transfected mammalian cell lines. rFVIII has proven to be a remarkable facsimile to pdFVIII with regard to biochemical and hemostatic properties, and has a similar pharmacokinetic profile with a half-life of roughly 12 hours. After almost 20 years of experience with rFVIII, there are clearly opportunities to extend recombinant DNA technology to further enhance replacement therapy. Bioengineering strategies have been directed at overcoming the inherent limitations of rFVIII biosynthesis and secretion, functional activity, half-life, and antigenicity/immunogenicity (Figure 1). Some of these strategies have already reached commercialization, several are in ongoing clinical trials, and a number of strategies are in advanced preclinical development. Although many of these strategies may be promising for replacement therapy and prophylaxis, others may find application in gene therapy studies.

Early on in the study of rFVIII expression, it was demonstrated that the B domain of FVIII, the equivalent of approximately 38% of the primary cDNA sequence, could be removed without loss of FVIII procoagulant activity. This significantly improved the yield of rFVIII due to markedly increased levels of mRNA and increased translation. The reduced size of the B-domain–deleted (BDD)-FVIII cDNA also facilitated packaging within certain viral vectors facilitating its adoption for gene therapy strategies. BDD-rFVIII remains the first and only modified human rFVIII molecule to come to commercialization. Other strategies to improve the efficiency of expression have included introduction of a truncated factor IX (FIX) intron 1,⁵  point mutations that reduce interactions with endoplasmic reticulum (ER) chaperones⁶  or improve the efficiency of ER-Golgi transport (through inclusion of a short B-domain segment within BDD-rFVIII⁷ ). Because these targeted modifications involve different steps within the secretion pathway, they can be combined in the same molecule providing an additive effect in producing substantial increases in the yield of FVIII in heterologous expression systems.⁷ 

Porcine BDD-rFVIII is in development for use in patients with inhibitors to human FVIII.⁸  Porcine BDD-rFVIII generates 10- to 14-fold higher expression than human BDD-rFVIII when expressed in baby hamster kidney cell lines.⁹  Subsequently, HEK-293 cell lines expressing porcine BDD-rFVIII demonstrated 36- to 225-fold higher expression than human rFVIII constructs. Furthermore, the higher protein production was not caused by significant increases in steady-state FVIII mRNA levels. This suggests a translational or posttranslational advantage for porcine rFVIII. Further study using porcine/human hybrids has localized certain protein sequences within porcine FVIII that confer this advantage, although additional studies will be necessary to characterize the mechanism.¹⁰  Bioengineering rFVIII for high-efficiency production could potentially reduce costs and increase the availability of concentrates for therapy.

Following activation of FVIII by thrombin, activated FVIII (FVIIIa) exists in a heterotrimeric form with its activity limited by additional proteolytic degradation and spontaneous decay through subunit dissociation. To address this limitation, an inactivation-resistant FVIII (IR8) was genetically engineered, which is not susceptible to dissociation of the A2 domain subunit and proteolytic inactivation by activated protein C.¹¹  This was achieved through modification of thrombin cleavage sites and the introduction of point mutations. IR8 exhibits an increased specific activity and prolonged cofactor function after thrombin activation in vitro and in vivo. Gale and colleagues¹²  introduced a disulfide bond (DSB) engineered between the A2 and A3 domains to stabilize FVIIIa by preventing A2 subunit dissociation following thrombin activation. A three-dimensional homology model of the FVIII A domains¹³  suggested that cysteines substituted at residues 664 and 1826 within FVIII would result in a DSB at the edge of the interface between the A1 and A3 domains, very near their solvent-exposed surfaces. This DSB-FVIII variant (C664-C1826) exhibited increased specific activity, prolonged cofactor activity following thrombin activation, and increased potency in whole-blood clotting assays.¹⁴  FVIII proteins with prolonged activity following thrombin activation have the potential to increase the efficacy of FVIII in plasma potentially prolonging the activity of FVIIIa, even when present in levels usually ineffective for hemostasis.

A major emphasis of current bioengineering efforts has been on half-life extension.¹⁵  Prolonging the half-life of FVIII could greatly reduce the frequency and dose of infusions, thereby improving the efficacy of prophylaxis through better compliance, as well as improve convenience and patient quality of life.¹⁶  The primary determinant of FVIII residence time in plasma is interaction with von Willebrand factor (vWF), which protects it from proteolysis and cellular uptake. Clearance of FVIII has only recently begun to be elucidated. FVIII is too large in molecular weight to be cleared by the kidneys. Cellular clearance (Figure 2), primarily in the liver, occurs through interaction with a family of low-density lipoprotein receptor-related proteins and heparan sulfate proteoglycan receptors, among others.¹⁷  Some of the half-life extension strategies under investigation include the following: sustained delivery through association of rFVIII with polyethylene glycol (PEG)ylated liposomes, chemical modification (eg, PEGylation, polysialylation), bioengineering rFVIII through mutagenesis at putative binding sites for clearance receptors, or the generation of fusion proteins (eg, with an Fc antibody fragment).¹⁵  It may be surprising that PEGylated rFVIII has not reached the clinic sooner. This has been used for a number of biologics and successfully transitioned to the clinic (eg, PEG-asparaginase for acute lymphoblastic leukemia). However, the primary advantage of PEGylation for a biologic is the incorporation of many water molecules within the hydrophilic PEG structures, functionally increasing the effective size of the conjugated protein above the filtration size of the kidney. This is of no particular advantage for FVIII, because it is already too large for kidney filtration. Any advantage to PEGylation of FVIII is likely through disruption of interaction with cellular clearance receptors. However, such chemical modification could be a disadvantage to FVIII if it interfered with key protein–protein interactions (eg, vWF, thrombin, and activated FIX [FIXa]). These potential hazards of chemical modification of FVIII are particularly problematic without the ability to target the sites where PEG polymers are conjugated to the protein.

Rostin et al¹⁸  published their experience with PEGylation of BDD-rFVIII. Random coupling of PEG at amino groups of lysines led a significant reduction in specific activity, as well as a significant proportion of molecules that could not bind vWF. Thus, recent strategies have focused on novel chemical conjugations that allow targeted PEGylation. Mei et al¹⁷  recently described their success by screening targeted PEGylated rFVIII mutants achieved through linkage to free surface-exposed cysteine residues introduced through mutagenesis. They identified variants that retained full procoagulant function and vWF binding in vitro, exhibited improved pharmacokinetics in hemophilic mice and rabbits, and prolonged efficacy in bleeding models in mice consistent with their enhanced half-life in vivo. Their results confirm that the site of PEGylation on FVIII is critical to PEGylation efficiency, preservation of procoagulant activity, and pharmacokinetic impact.

Polysialic acid (PSA) modification of therapeutic proteins is an alternative strategy to PEGylation for increasing the size of a protein. PSAs are linear, hydrophilic polymers of N-acetylneuraminic acid that occur abundantly on the surface of many cells and proteins. They can be conjugated to therapeutic proteins and potentially alter their pharmacokinetics, including prolonging the half-life.¹⁹  PSA also produces a “watery cloud” around the therapeutic molecule protecting it from immune-mediating cells, proteolytic enzymes, and clearance receptors. The use of PSA has already been successfully applied to several therapeutic proteins.²⁰  Some of the advantages that have been demonstrated with PSA technology include reduced immunogenicity and antigenicity, and preservation of function with increased stability. Furthermore, PSA, unlike PEG, are naturally occurring and biodegradable, which offers a potential advantage when administering large doses of a biologic therapy over a prolonged period of time. PSA modification is now being explored through direct conjugation of rFVIII, FIX, FVIIa, and vWF.

Formulation with PEGylated liposomes is an established process that has been used to extend the half-life of a broad range of therapeutic proteins, including FVIII.²¹  For this approach, the therapeutic protein is reconstituted with PEGylated liposomes as a carrier. Because liposomes are typically cleared very quickly from the circulation, the addition of PEG can extend the circulatory half-life of the liposomes considerably. This approach can effectively modify pharmacokinetic and pharmacodynamic properties of proteins and has been utilized to develop a potentially longer acting FVIII. With this strategy, rFVIII molecules remain unmodified, so there is no loss of normal protein–protein interactions and functional activities. In preclinical trials within a hemophilia A mouse model, prophylactic infusion of rFVIII reconstituted with PEGylated liposomes (rFVIII-PEG-Lip) prolonged some pharmacokinetic parameters, compared with standard rFVIII and correlated with an enhanced hemostatic efficacy.²²  In addition, rFVIII-PEG-Lip has been examined in patients with severe hemophilia A in a blinded, controlled, crossover, multicenter trial.¹⁶  A single prophylactic infusion of rFVIII-PEG-Lip resulted in a longer bleed-free interval, compared with standard rFVIII. The rFVIII-PEG-Lip formulation was well tolerated, and no significant adverse events were reported during the trial. These findings generated great interest, and a subsequent, double-blind, randomized, crossover phase I trial was conducted to compare the pharmacokinetics of a single infusion of rFVIII-PEG-Lip with that of rFVIII in 26 men with severe hemophilia A.²³  However, rFVIII-PEG-Lip and standard rFVIII demonstrated similar pharmacokinetic parameters. Additional preclinical studies in a hemophilia A mouse model suggested that rFVIII-PEG-Lip results in prolonged apparent FVIII activity beyond what would be expected by the plasma pharmacokinetics when whole blood clotting was assayed by rotational thromboelastography.²⁴  Furthermore, rFVIII-PEG-Lip increased P-selectin surface expression on platelets in response to collagen, and the enhanced procoagulant activity was retained until ultracentrifugation of the plasma, suggesting it may be acting through sensitization of platelets and the generation of procoagulant microparticles.²⁵  Nevertheless, a phase II clinical trial was stopped midstage when an interim analysis indicated that the trial would not meet its efficacy endpoint.

Another line of investigation involves developing FVIII with reduced antigenicity/immunogenicity. The most sophisticated constructions are human/porcine hybrid BDD-rFVIII molecules that have reduced antigenicity within human inhibitor plasmas,²⁶  and have guided the design of bioengineered variants with reduced immunogenicity in a hemophilia A mouse model without any apparent loss of function.²⁷  In another strategy, BDD-rFVIII has been produced in the human embryonic kidney cell line, HEK293F cells.²⁸  This was done to produce a human pattern of posttranslational modifications of FVIII, such as glycosylation, to reduce the immunogenic profile, compared with existing rFVIII produced in hamster cells. Further clinical studies will need to determine if there is a significant advantage of these types of FVIII variants in previously untreated patients with hemophilia.

Purified pdFIX has been available since 1992, followed by recombinant FIX (rFIX) in 1998, produced in mammalian cell lines similar to rFVIII. Although, rFIX also shares nearly identical hemostatic properties with pdFIX, there are differences in posttranslation modifications of rFIX that result in altered pharmacokinetics. Specifically, rFIX exhibits an ∼ 30% reduced recovery in plasma at equivalent dosing to pdFIX, requiring dose modification for treatment and prophylactic regimens. However, the plasma half-life for both pdFIX and rFIX is similar at roughly 16 hours.

An innovative expression technology has been used to enhance the expression of rFIX in stable cell lines.²⁹  This strategy involves flanking rFIX cDNA with the transcriptional control regions from Chinese hamster elongation factor 1α, a highly expressed gene within Chinese hamster ovary (CHO) cells. Within this system, the expression levels of rFIX are increased 10-fold over traditional stably transfected CHO cells. Such technology could be exploited for efficient production of rFIX, or any other recombinant protein, and hopefully facilitate reduced costs.

Transgenic animals have shown promise as an efficient source of producing abundant recombinant proteins.³⁰  The yields from such systems could improve the availability of recombinant clotting factors worldwide or serve as an efficient source material for chemical modification to alter pharmacokinetic or biochemical properties, or to pursue alternative delivery systems (eg, oral administration). Recombinant coagulation protein production requires that the cell culture bioreactor systems can replicate the complex biochemistry of their plasma-derived counterparts and is influenced by the cell density within the bioreactor system. The mammary gland of livestock has both the cellular machinery and high cell density to produce recombinant proteins at two orders of magnitude or greater than traditional cell culture bioreactor systems. The first approved transgenic recombinant product was recombinant antithrombin produced in the milk of transgenic goats.³¹  However, in contrast to ruminants, the pig mammary gland has demonstrated the capacity to carry out the posttranslational modifications (eg, glycosylation, sulfation) necessary to enable high efficiency production of recombinant proteins with the biologic activity and pharmacokinetic properties needed to be considered as a therapeutic for hemophilia. Recombinant human FIX has been produced in the pig mammary gland at very high yields (100 IU/mL), compared with plasma source material (1 IU/mL) or CHO cell bioreactors (2 IU/mL), such that it has been estimated that the milk from 60 pigs (∼ 12,000 L/year) could supply the entire amount of rFIX needed for prophylaxis in the United States.³² 

Several bioengineering strategies have generated rFIX variants with increased potency. These include point mutations, such as variant FIX-R338A, which exhibits 3- to 7-fold higher specific activity in vitro³³  and in vivo.³⁴  Recently, the importance of this particular residue was highlighted by the identification of a family with a potent X-linked thrombophilia who were identified as carrying a R338L missense mutation of FIX.³⁵  The plasma FIX activity of the proband was 800% of normal, and an rFIX-R338L exhibited a 5- to 10-fold higher specific activity in vitro. Hopfner et al³⁶  used insights from the structural and functional homology between FIXa and activated factor X (FXa) to bioengineer a recombinant FIX-FX hybrid that exhibited a catalytic efficiency 130-fold that of the wild-type rFIX. Such variants could considerably reduce the infusion requirements for hemostatic efficacy.

Strategies to extend the half-life of FIX include direct, targeted modification with PEGylation and fusion protein technology, which links FIX to another protein with a much longer plasma half-life. rFIX was fused to the constant region (Fc) of immunoglobulin G (rFIXFc).³⁷  The presence of the Fc portion protects the fusion protein from catabolism through interaction with the neonatal Fc receptor (FcRn). The Fc domain binds FcRn at acidic pH (< 6.5), but not at physiologic pH (7.4). Fc-containing proteins that are internalized by endothelial cells bind to FcRn present in the acidified endosome in a pH-dependent manner and are then recycled back to the cell surface where they are released back into plasma at physiologic pH, rather than targeted for degradation in the lysosome. Preclinical data with rFIXFc demonstrated a 3- to 4-fold longer terminal half-life in mice, rats, and cynomolgus monkeys. The functional impact was demonstrated as the whole-blood clotting time was corrected through 144 hours for rFIXFc, compared with 72 hours for unconjugated rFIX in the hemophilia B mouse model. Similar results were obtained in two hemophilia B dogs. In a second strategy, rFIX was fused to albumin (rFIX-FP) via a cleavable peptide linker.³⁸  This allows for release of rFIX from albumin on activation. rFIX-FP demonstrated significantly prolonged pharmacokinetics in rats, rabbits, and hemophilia B mice with effective hemostasis. Both of these fusion protein strategies are presently being evaluated in clinical trials.

Recombinant activated FVII (rFVIIa) has proved to be a safe and effective therapeutic for the management of bleeding in hemophilia patients with inhibitors. It has both tissue factor-dependent and independent mechanisms of action. However, its short plasma half-life requires a short interval for follow-up dosing and limits its application in prophylaxis. In addition, clinical studies have suggested that higher initial doses may result in a more rapid onset of hemostasis,^39,40  and the time from initiation of bleeding to initiation of treatment may have a significant impact on efficacy.⁴¹  Accordingly, several bioengineering strategies have been implemented in an attempt to improve rFVIIa functionality. Some of these include glycoPEGylation,⁴²  targeted PEGylation to specific residues of FVII, formulation with PEG-Lip,⁴³  and fusion of rFVIIa with albumin,⁴⁴  all directed at half-life extension. Other strategies have attempted to increase the potency and rate of onset of action of rFVIIa through directed molecular evolution and rational design.^45,46  Several of these approaches have already moved to the clinical testing phase.

Nonintravenous (ie, oral, intratracheal, subcutaneous, and intramuscular) delivery of hemostatic agents for hemophilia would liberate patients from the challenges of regular direct venous access and the complications of central venous catheters. However, the very low bioavailability observed with such nonintravenous delivery of already expensive clotting factor concentrates remains a barrier. In addition, some routes of administration could actually increase clotting factor immunogenicity. Alternative delivery strategies could be realized if the production costs of concentrates could be dramatically reduced. Recombinant clotting factor expression within transgenic animals and plants may be able to reach such production scales. Additional bioengineering strategies could also improve bioavailability via the enteric route. Transgenic plants expressing rFIX fused to a transmucosal carrier protein have been prepared.⁴⁷  The rFIX remains encapsulated within the plant cells, protecting it from enteric degradation. Interestingly, hemophilia B mice that were fed this bioencapsulated transgenic rFIX plant material were also protected from inhibitor formation and anaphylaxis to intravenous challenge with rFIX, suggesting such oral delivery strategies could be a strategy to avoid another major complication of protein replacement therapy.

Research toward novel therapeutics for hemophilia is not limited to bioengineering recombinant coagulation proteins. Alternative protein and nonprotein agents are showing promise in preclinical development. Kopecky et al⁴⁸  synthetically designed peptide sequences that were screened and tested for their ability to bind to inhibitory antibodies effectively neutralizing anti-FVIII antibodies in vitro. The promising peptides typically have sequences that are similar to FVIII. However, owing to their short half-life, strategies such as PEGylation may be required to extend their circulation time.

Fucoidans, known also as nonanticoagulant-sulfated polysaccharides (NASP), are heparin-like molecules that at certain concentrations exhibit hemostatic properties rather than anticoagulant activity. They are believed to act through blockade of tissue factor pathway inhibitor, effectively removing the brakes from coagulation. NASP can accelerate the clotting times of plasma from hemophilia patients, and improve hemostasis when administered subcutaneously to hemophilic mice⁴⁹  and dogs.⁵⁰  Most notably, NASP improved hemostasis when administered orally to severe hemophilia A dogs.⁵⁰ 

Aptamers are single-stranded nucleic acids that can directly inhibit function by folding into a specific three-dimensional structure with a high affinity for the target. They can theoretically be generated to bind to any protein of interest. This technology has been explored in clinical trials of anticancer and antiviral therapeutics.⁵¹  Moreover, opportunities also exist for therapeutic aptamers as antithrombotic and hemostatic agents. Aptamers that target tissue factor pathway inhibitor have shown similar hemostatic effects in vitro as NASP.⁵² 

Finally, missense mutations leading to premature stop codons in FVIII or FIX occur in approximately 10% to 15% of persons with hemophilia A or B, respectively. Small molecules have been developed that can facilitate translational readthrough of premature stop codons. A lead candidate has been tested in transgenic mice engineered with a nonsense FIX mutation (R338X) and resulted in detectable plasma levels (3-5 ng/mL) in up to 20% of the mice.⁵³  A phase II clinical trial is underway in hemophilia patients known to carry premature stop codons.

Although a cure for hemophilia has not yet been realized, there is a tremendous pipeline of novel therapeutic agents. Promising new bioengineering strategies are being driven by detailed structure and function analysis of coagulation proteins complimented by high-resolution crystal structures. A broad range of species now serves as preclinical models to guide lead candidate selection. In addition, alternative therapeutics with novel mechanisms of action may dramatically alter therapeutic paradigms. The best ideas from the bench have moved from preclinical studies to clinical trials (Figure 4). We can look forward to many of these strategies successfully overcoming some of the remaining challenges that remain for the treatment of hemophilia.

Conflict-of-interest disclosure: The author has received honoraria from Baxter BioScience, Inspiration Biopharmaceuticals, CSL Behring, and Novo Nordisk, and research funding from Inspiration Biopharmaceuticals. He has also been a consultant for Baxter BioScience. He has a membership on an entity's Board of Directors or advisory committees for Baxter BioScience and Novo Nordisk.

Off-label drug use: None disclosed.

Steven W. Pipe, MD, Department of Pediatrics and Communicable Diseases, Hemophilia and Coagulation Disorders Program, University of Michigan, MPB D4202, 1500 E. Medical Center Dr., Ann Arbor, MI 48109; Phone: (734) 647-2893; Fax: (734) 615 (0464); e-mail: ummdswp@med.umich.edu